Short Communication: Geometric morphometric analysis of leaf venation in four Shorea species for identification using Digital Image Processing

Abstract

Abstract. Ariawan I, Herdiyeni Y, Siregar IZ. 2020. Short Communication: Geometric morphometric analysis of leaves venation in Shorea spp. for identification using Digital Image Processing. Biodiversitas 21: 3303-3309. Shorea is one of the genera of the Dipterocarpaceae family which consists of more than 190 species. Massive exploitation of forests has threatened the sustainability of Shorea in nature. A total of 156 species has been listed on the IUCN (International Union for Conservation of Nature) red list. From the 156 species, 59.6% are in the critically endangered category, so urgent conservation is needed. However, during collection of Shorea at the seedling phase for conservation purposes, it is often difficult to distinguish among them that can cause errors in their collection process. To avoid these errors, identification needs to be done, usually based on plant leaf and flower morphology. Leaves are easier because they have the main features that distinguish each plant species, one of which is the venation structure. Geometric morphometric techniques are a modern approach recognized as useful for the identification of species in many plants. Geometric morphometrics analyzes the position of the venation point using coordinate geometry values. This research was aimed to extract venation features of Shorea leaves using a geometric morphometric approach. The extraction process result in some features, such as straightness, different angle, length ratio, scale projection, and secondary nerves. On extracted features, an analysis was then performed to find out the best features in classifying species of Shorea spp. The results of this study indicated that the geometric morphometric approach could extract the value of the features of straightness, different angle, length ratio, scale projection, and secondary nerves. The secondary nerve feature is the best feature because it can distinguish between fourcommonly planted species of Shorea spp. (S. acuminata, S. leprosula, S. ovalis, and S. selanica). By using the support vector machine classification technique to identify species of Shorea spp., the classification results obtained an average accuracy of 84.46%.

Klasifikasi Tiga Genus Ikan Karang Menggunakan Convolution Neural Network

“…Salah satu metode analisis tersebut adalah klasifikasi. Klasifikasi citra merupakan metode yang digunakan untuk mengelompokkan informasi yang diperoleh dari citra yang diamati (Ariawan et al, 2020 ;Kurniawati et al, 2020;Ariawan et al, 2022). Convolutional Neural Network (CNN) merupakan salah satu metode klasifikasi yang dapat digunakan untuk pengelompokkan citra (Putra et al, 2016).…”

“…Jika lapisan CONV, e, mempunyai fitur inputan sebesar dari output lapisan sebelumnya, dengan masing-masing berukuran , maka keluaran dari e yaitu sebesar fitur dengan ukuran . Fitur ke -i pada lapisan e, , dapat dihitung menggunakan persamaan berikut (Rokhana et al, 2019): ................. (2) adalah bias dari matriks, adalah filter untuk fitur ke-j pada lapisan ( ) dan fitur ke-i pada lapisan e, ukuran filter , maka D dapat dituliskan pada persamaan berikut (Rokhana et al, 2019): ............... (3) Menghitung ukuran fitur keluaran lapisan e dapat menggunakan persamaan berikut (Rokhana et al, 2019): .................................... (4) .................................... (5) CNN mempunyai fungsi aktivasi ReLU untuk mempercepat konvergensi dari proses pelatihan dengan meningkatkan kekurangan dari jaringan menggunakan koefisien gain ( ), sehingga korelasi tersebut dapat disederhanakan menggunakan persamaan berikut (Ahmad et al, 2019): .................................... (6) .............................. (7).…”

Geometry Feature Extraction of Shorea Leaf Venation Based on Digital Image and Classification Using Random Forest

“…In the field, species identification is usually conducted using leaves [7], [8], [9]. These plants morphological traits are easy to locate in the forest floors that exhibit several distinct characteristics [10].…”

“…Properties applied to differentiate fruit trees are growth, include angle (Ang), length (Len) and distance (Dis) [16] while to classify medicinal plants are densities of leaf vein, branching and ending points were employed [15]. However, to distinguish Shorea species, four attributes including, scale projection (Spr), angle difference (Adi), straightness (Str), and length ratio (Lra), were involved [9]. This research generally applied the feature angle, length, distance, scale projection, angle difference, straightness, and length ratio, as well as densities of leaf vein, branching and ending points, in order to identify Shorea species.…”

Morphological evaluation and determination keys of 21 citrus genotypes at seedling stage

“…Previous study proved that leaf venation structure help to distinguish 4 common species of Shorea spp. with accuracy of more than 80% (Ariawan et al 2020). Leaf shape is successful to solve some misunderstandings in the taxonomy of Sagittaria trifolia (Huang and Liu 2014).…”